PROJECT SUMMARY/ABSTRACT Opioid overdose, occurring as part of opioid mono-substance use (MSU) and, more intensely, among those with poly-substance use (PSU), has reached an epidemic level in the United States, and is reflective of other high-risk behaviors among those with PSU. Accordingly, there is an urgent public health need for improved treatment and prevention strategies for these populations, as guided by findings on the etiology and nature of both laboratory-assessed and clinical harm-avoidance deficits. Our purpose is to evaluate the nature and significance of laboratory-assessed harm avoidance deficits in order to identify potential treatment targets (mechanistic outcomes) that may underlie devastating, clinical harm-avoidance deficits, characterized by high risk for blood-borne viruses, criminality, and repeated overdoses among those with opioid PSU relative to MSU. In the animal model, we examine whether laboratory-assessed harm avoidance deficits (assessed via drug self-administration punishment and subsequent operant avoidance paradigms) arise from pre-existing inhibitory control deficits, drug use, or the combination of these factors, specifically evaluating differences in harm-avoidance capacity in mono (heroin or cocaine) vs. poly (heroin + cocaine) substance use models. Our subsequent work in humans will use two fear-based (acquisition of learned fear association, operant avoidance of this association) and two monetary-based (Iowa Gambling Task and monetary choice) laboratory measures of harm avoidance deficits to compare the severity of laboratory harm avoidance deficits in outpatients with opioid MSU or opioid+stimulant PSU. We will use these measures to predict and understand the nature of clinical harm-avoidance deficits in this sample of men and women. Close analogue procedures have been built into the animal and human studies, with the human studies providing a quasi-experimental or associative replication of the animal paradigms, i.e., allowing the tightly controlled experimental evidence (allowing causal conclusions) from the animal phase to aid the design and interpretation of less controlled, predictive models in the human stage.